Integration: Mem0

Add persistent, user-specific memory to your Haystack agents and pipelines with Mem0

Authors

deepset

Mem0

Table of Contents

• Overview
• Installation
• Usage
  • Standalone Memory Operations
  • Using Mem0 with a Haystack Agent
• License

Overview

Mem0 (pronounced “mem-zero”) provides a universal memory layer for AI agents and assistants. It enables your Haystack applications to remember user preferences, adapt to individual needs, and continuously learn from past interactions — making AI conversations truly personalized.

This integration is part of the haystack-experimental package and provides the Mem0MemoryStore, which acts as a persistent memory backend for Haystack Agents. Instead of relying solely on conversation history or static document stores, agents can use Mem0 to store and retrieve user-specific memories across sessions.

Key Features

• Persistent Memory: Store and retrieve user-specific memories that persist across sessions
• Multi-Level Scoping: Organize memories by User, Session, or Agent scope
• Intelligent Extraction: Automatically extracts relevant facts from conversations — no need to store entire transcripts
• Seamless Agent Integration: Works natively with Haystack’s Agent component for context-aware responses
• Flexible Deployment: Use the managed Mem0 Platform or self-host with your own infrastructure

More info about Mem0:

• Mem0 Website
• Mem0 Documentation
• Mem0 Platform (Managed)
• Mem0 GitHub

Installation

pip install haystack-ai haystack-experimental mem0ai

Environment Variables

Set the following environment variable to use the Mem0 Platform:

export MEM0_API_KEY="your-mem0-api-key"

You can obtain an API key by signing up at app.mem0.ai.

Usage

Components

This integration introduces one component:

• Mem0MemoryStore: A memory store that uses Mem0 as the backend for storing and retrieving user-specific memories. It can be used standalone or plugged into a Haystack Agent.

Standalone Memory Operations

You can use Mem0MemoryStore directly to add and search memories:

import os
from haystack.dataclasses import ChatMessage
from haystack_experimental.memory_stores.mem0 import Mem0MemoryStore

os.environ["MEM0_API_KEY"] = "your-mem0-api-key"

# Initialize the memory store
memory = Mem0MemoryStore()

# Add memories from a conversation
messages = [
    ChatMessage.from_user("I'm a vegetarian and I love Italian food."),
]
memory.add_memories(messages=messages, user_id="alice")

# Later, retrieve relevant memories
query = "What kind of food do I like?"
memories = memory.search_memories(query=query, user_id="alice")

print(memories)
# Returns memories related to Alice's food preferences

Using Mem0 with a Haystack Agent

The primary use case for Mem0MemoryStore is to provide personalized context to a Haystack Agent. The agent automatically retrieves relevant memories before generating a response, enabling context-aware conversations:

import os
from haystack.components.generators.chat import OpenAIChatGenerator
from haystack.dataclasses import ChatMessage
from haystack_experimental.components.agents import Agent
from haystack_experimental.memory_stores.mem0 import Mem0MemoryStore

os.environ["MEM0_API_KEY"] = "your-mem0-api-key"
os.environ["OPENAI_API_KEY"] = "your-openai-api-key"

# Initialize components
memory_store = Mem0MemoryStore()
generator = OpenAIChatGenerator(model="gpt-4o")

# Create an agent with memory
agent = Agent(
    generator=generator,
    memory_store=memory_store,
    system_prompt="You are a helpful assistant that remembers user preferences.",
)

# First interaction — the agent learns about the user
response = agent.run(
    query="I prefer dark mode and use vim keybindings.",
    memory_store_kwargs={"user_id": "user_123"},
)
print(response["replies"][0].content)

# Later interaction — the agent recalls the user's preferences
response = agent.run(
    query="Can you recommend an IDE setup for me?",
    memory_store_kwargs={"user_id": "user_123"},
)
print(response["replies"][0].content)
# The agent will reference the user's preference for dark mode and vim keybindings

Note: The Mem0MemoryStore is currently part of the haystack-experimental package and is labeled as experimental. The API may change in future releases. Always refer to the haystack-experimental repository for the latest updates.

License

haystack-experimental is distributed under the terms of the Apache-2.0 license.